Current size predicted timing jitter better than a shared gamma clock with independent units in ten mouse cortical connections

We tested whether functional release-unit estimates predict timing jitter with one common release clock. We joined exact source labels for five prefrontal and five sensory mouse cortical connections. The model used independent units, their measured release probabilities, and a gamma clock of shape two. With each test date absent from the fit, one fitted time scale gave mean squared log error 0.225. Area means gave 0.409. The 45% point gain had a descriptive date resampling range from 95% more error to 82% less error. A rule based on current size gave error 0.094, or 58% less than the release model. After a fit on the ten records, this rule gave error 0.195 in 117 other connections from 79 other dates. Area means fitted on the same ten records gave 0.403. This added test was selected after the first comparison. Joint date resampling with refits gave a gain range of 22% to 79%. An added common-offset clock did not improve mean delay forecasts. The result does not separate vesicle timing from onset-fit error in small currents. Functional release-unit estimates are not anatomical synapse counts. The package provides checked source matches, code, first-release moments, figures, forecasts, and repeat instructions.

synaptic timing · quantal release · release units · first release · cortical circuits · prefrontal cortex · somatosensory cortex · prediction · connectome models